An Emburse survey earlier this year of more than 1,000 employees and executives showed that around a quarter of employees admitted to trying to pass off a personal expense as a business expense.

But Kasin says when he speaks to C-suite finance executives, they frequently are unaware of the potential costs of expense fraud. "What I have found out in these meetings is that they understand the aspect of what expense fraud is, but no one understands the depth and breadth of what we're talking about," he says.

Technology, of course, plays a key role in detecting and preventing expense fraud, and advancements with AI and machine learning as well as opportunities to tap into more wholistic travel and expense platforms stand to reduce those numbers. Travel and expense managers, however, will need to be strategic in their approaches to combatting fraud, says Pretium global director of travel and expense Maria Chevalier.

"Fraud is not limited to employees," she says. "All these groups are getting new and more sophisticated ways to commit fraud with a company's travel and credit card programme."

The payment line of defence

There's no better way to reduce fraud costs than to prevent fraudulent charges from occurring in the first place, and payment technology can provide controls to keep purchases in policy and reject illicit purchases outright. The traditional payment and expense management approach, which is still in place for "90 per cent of the market," is not set up for that, says Center cofounder and SVP of product Nik Singh.

"You have an expense management system that is being populated by and interacting with personal cards that people are manually entering in data or people taking a snap of a receipt that may or may not be legible or disparate corporate cards that don’t connect," he says. "What that has led to is inaccurate data, which leads to an inability to understand where spend is going, and poor controls."

Solving for that has been a key driver of recent deepening of connections between fintech companies and travel. Center has partnered with Spotnana, with which it shares investor DNA, to create an integrated travel and expense offering using its card product. In recent weeks, Navan announced a partnership with Brex to launch a payment solution integrated into its platform. Payment and expense platform Mesh launched its own travel management platform last year.

With an integrated solution, companies have the capability not only to issue payment tools to all travellers but also to put controls around those tools, says Singh. Not only might that include spending limits or blocking certain merchant categories – not allowing spending at a casino, for example – but it also gives the ability to tailor payment to employee "personas." A business traveller, for example, would need the ability to pay for hotels and meals, but the company might want to block procurement purchases on that tool, while a professional services team would need access to pay for fuels and meals but not to book air.

"This concept of being able to limit the types of spend that apply to cards is a huge innovation that drives this pre-spend control process," says Singh.

Chevalier said companies vary on their requirements and strategies which can drive the different benefits and impact when comparing the all-in-one as compared to the independent providers in terms of assisting in controlling expenses and mitigating risk.

There are other options to have controls at the point of payment. SAP Concur's Verify auditing service can issue alerts to travellers throughout their trip to head off potential out-of-policy spending, such as if a dinner comes in well over the cost of a daily per diem allowance, says Concur head of market strategy Christopher Juneau.

"As the traveller is spending while they travel, you're giving travellers alerts to add attendees, make sure you have the receipt or know something is potentially out of policy," he says. Doing that is delivering a better experience for the traveller but reminding them that the system is watching them as well."

As virtual card adoption increases, that will bring more spending controls not only in terms of what employees spend but also in reducing the risk of stolen cards, says Emburse chief technology officer Ken Ringdahl. Payment providers, in the meantime, have been honing their own technology in terms of detecting misuse. Even though the liability generally falls to the credit card companies in such situations, that doesn't mean it shouldn't be a concern for the travel management side, says Chevalier.

"While the credit card companies may own most of the financial risk of the various fraud categories, it is critical that this is managed collaboratively with all your travel suppliers to assist in the identification of potential fraud but also that there are processes to prevent it," she says. "We need to prevent fraud while protecting their data and privacy."

Beyond that, companies should consider the resources involved in trying to track down and report such misuse, says Ringdahl. The cost of fraud is more than the "hard dollar" amounts; there's also the soft-dollar cost of employee time spent tracking down and correcting it, Singh adds.

A digital eye on receipts

Receipt image capture technology is nothing new, but some suppliers are refining it to capture increasingly granular detail enhanced by AI and machine learning, which is making it more powerful in countering fraud.

At the heart of receipt imaging tools is optical character recognition technology, which can convert images of text into text that is readable by machines. As receipts come in a variety of formats and texts, including hand-written, it has taken time for the technology to improve its efficiency.

Emburse's Rindgahl says the company's "intelligent" OCR technology, which can pull out items on a line-by-line basis from receipts, enables customers to "set pretty complex policies" supported by receipt imaging. Policies that do not allow reimbursement for purchases of alcohol is a popular one, with the imaging technology able to flag line items of alcohol and subtract them from the total, he says.

Not all claims of OCR are equal, however, says Philipp Schloter, president and CEO of expense management provider Abukai, which is centered around its receipt capture technology. If it is not able properly to capture amounts and categorise them, requiring users to manually edit them or fill in the blanks, it is "back to almost filling out the expense report manually."

Schloter says Abukai has the capability to "generate more than 60 fields of data" from a receipt and audit it against profiles and policy.

"Before you can detect fraud, you need data," says Schloter. "You need something that's easy and where you're getting additional data capture of what that purchase was for."

Juneau says receipt imaging improves data, which helps with analytics and fraud management. Among customers that use Concur's ExpenseIt receipt imaging technology – which recently has added enhanced capabilities via AI to help with hotel folio itemisation – that the number of expense reports that are returned to the end user goes down 90 per cent, he says.

AI patrols

Most of the major travel management companies are looking at introducing AI into their workflows, if they have not done so already. Cornerstone’s Mat Orrego says, "We have run a pilot with one company where we were able to create a more flexible policy using a chat bot and an approval process within that."

A large part of expense auditing to prevent fraud is "looking for anomalies or patterns in the data," which "humans aren't great at looking for," says Juneau. A human auditor going through piles of company expense reports might miss the employee who somehow always has paid-in-cash lunches always $1 below the receipt threshold or the same business dinner expensed by three different employees attending it. Expense suppliers already are using AI and machine learning to identify such patterns, and they are honing those capabilities as generative AI technology strengthens.

"Today, AI is helping with the actual capture of the data itself," says Singh. "You take a corporate card transaction, clean it up and present that in real time, giving that end-user individual much more data than they had in the past."

As an example, AI can analyse a meal transaction and tell that it is a business meal based on location, he says. AI also can take Level 2 and Level 3 data from transactions and itemise purchases.

Effective fraud prevention requires integration of multiple data sources, which is the approach of BizTravelSolutions with its Evita (Expense Verification Information Through Analytics) tool, says Kasin. The tool can take a CSV file and travel policies and employ algorithms to analyse and verify the data, he says.

Where generative AI will likely move expense analysis is to act more like a "financial agent" in guiding actions based on its analysis, says Singh.

"Think of it as a financial partner who is helping to sift through the data," Singh says. "Instead of you having to dig in and derive the insights, it's going to start to tell you this is where we see oddities or problems, and here are our recommendations for how to go into the system and change things."

That still doesn't fully eliminate the human element from the fraud prevention equation, but it makes them more tactical, responding to anomalies rather than looking for them. Across the organisation, that can make for smoother implementation of policies, both from the traveller and management perspective.

"Historically, organisations have had to make a tradeoff between user experience and efficiency with compliance," says Juneau. "You set your audit rules, workflows and level approvals, and those things reduce the experience of the traveller and the efficiency of the process. We believe that with generative AI, we have the chance to deliver a better user experience but also drive greater compliance."